Hexameric ring structure of the N-terminal domain of Mycobacterium tuberculosis DnaB Tapan Biswas and Oleg V. Tsodikov

Department of Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor, MI, USA

Keywords Hexameric DnaB helicase unwinds the DNA double helix during replica- crystal structure; DnaB helicase; ; tion of genetic material in . DnaB is an essential bacterial ; replication; tuberculosis therefore, it is an important potential target for antibacterial drug disco- very. We report a crystal structure of the N-terminal region of DnaB from Correspondence ˚ O. V. Tsodikov, Department of Medicinal the pathogen Mycobacterium tuberculosis (MtDnaBn), determined at 2.0 A Chemistry, College of Pharmacy, University resolution. This structure provides atomic resolution details of formation of Michigan, Ann Arbor, MI 48109-1065, of the hexameric ring of DnaB by two distinct interfaces. An extensive USA hydrophobic interface stabilizes a dimer of MtDnaBn by forming a four- Fax: +1 734 647 8430 helix bundle. The other, less extensive, interface is formed between the Tel: +1 734 936 2676 dimers, connecting three of them into a hexameric ring. On the basis of E-mail: [email protected] crystal packing interactions between MtDnaBn rings, we suggest a model (Received 19 March 2008, revised 8 April of a helicase–primase complex that explains previously observed effects of 2008, accepted 11 April 2008) DnaB mutations on DNA priming. doi:10.1111/j.1742-4658.2008.06460.x

The DnaB gene encoding DnaB helicase was originally required for formation of the functional helicase identified in a region of the bacterial genome, mutations [11,12] and for binding DnaG [13], and a C-terminal in which caused a defect in DNA replication at elevated RecA-like domain, responsible for DNA binding and temperatures in vivo [1]. DnaB is an ATP-dependent ATP hydrolysis (Fig. 1A). Six monomers of DnaB helicase that unwinds parental duplex DNA during form a two-tiered ring, as observed in recent modest- replication [2–5], thereby allowing the two strands to be resolution structures of DnaB from the thermophilic copied. A hexamer of DnaB encircles the lagging strand Geobacillus stearothermophilus (Gs) [14], and in a low- and moves on it in the 5¢fi3¢ direction, thereby split- resolution structure of a DnaB homolog G40P from ting the two complementary DNA strands [6]. A hexa- Bacillus subtilis phage SPP1 [12], reported while this meric DnaB helicase alone is not very processive [7], article was in preparation. One tier is formed by whereas it becomes highly processive when bound to DnaBn, which has three-fold symmetry. The other tier other replication [8]. During DNA replication, is formed by the RecA-like ATPase domains in DnaB is associated with multiple copies of DNA prim- approximate six-fold symmetry. The hexameric compo- ase DnaG, which catalyzes the formation of RNA prim- sition of DnaB is shared by other representatives of ers for the synthesis of on the the same superfamily of replicative , such as lagging strand. The DnaB–DnaG complex plays a key bacteriophage T7 gene 4 protein [15], although sig- role in coordinating the mechanistically different contin- nificant differences in the sequence and domain orga- uous and discontinuous DNA syntheses on the leading nization exist among the family members. and the lagging strands, respectively [9]. DnaB helicase is conserved in the eubacterial king- A monomer of DnaB consists of two domains [10]: dom, and is distinct from its functional relatives an N-terminal helical domain (DnaBn), shown to be in higher eukaryotes. Therefore, DnaB is a very

Abbreviations DnaBn, N-terminal helical domain of DnaB helicase; DnaGc, C-terminal domain of DnaG; Gs, Geobacillus stearothermophilus; HhH, helix– hairpin–helix (HhH); Mt, Mycobacterium tuberculosis.

3064 FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS T. Biswas and O. V. Tsodikov Mycobacterium tuberculosis DnaB helicase

Fig. 1. (A) Domain organization of MtDnaB. The numbers indicate approximate domain boundaries. The C-terminal domain of DnaB in M. tuberculosis and several other mycobacteria is disrupted by an intein, as indicated. (B) Sequence alignment of DnaBn. Sequences of DnaBn from several bacterial species (Mt, M. tuberculosis; Ml, Mycobacterium leprae; Ms, Mycobacterium smegmatis; Gs, G. stearother- mophilus; Ec, E. coli) are aligned with the secondary structures of MtDnaBn (above; this study) and GsDnaBn [14]. The residues in the HhH–HhH interface of MtDnaBn hexamer are indicated by black bars, and those in the dimer–dimer interface are indicated by black circles. The surface residues of MtDnaB that form the putative DnaG-binding pocket (this study) are indicated by open red circles, and those of GsDnaB that contact DnaG in the GsDnaB–GsDnaG structure [14] are indicated by filled red circles. (C) SDS ⁄ PAGE gel of purified MtDnaB(21–197) and MtDnaB(21–134) proteins. (D) A gel filtration chromatogram demonstrating that MtDnaB(21–197) elutes as a dimer and MtDnaB(21–134) as a monomer. attractive target for antibacterial therapy. The C-ter- We report a 2.0 A˚resolution structure of the hexa- minal catalytic domain of DnaB is very highly con- meric ring DnaBn from the pathogen Mycobacte- served, which points at the conservation of the rium tuberculosis (MtDnaBn). On the basis of the ATPase function in this helicase superfamily. DnaBn organization of the MtDnaBn rings in the crystal is less conserved, and its binding partner, the C-ter- lattice, we propose a model of the mycobacterial minal domain of primase DnaG (DnaGc), is very DnaB–DnaG complex. weakly conserved. This divergence is intriguing, because it must be related to documented differences Results and Discussion in intrinsic stabilities of DnaB–DnaG complexes between Escherichia coli and G. stearothermophilus Structure of the hexameric ring of MtDnaBn [13,16–18]. This, in turn, implies divergent coordina- tion of unwinding with priming at the replication Bacterial DnaB helicase consists of two domains: the fork in different bacteria. moderately conserved, entirely helical DnaBn ( 24%

FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS 3065 Mycobacterium tuberculosis DnaB helicase T. Biswas and O. V. Tsodikov identity) (Fig. 1B) and the highly conserved C-terminal one of the projections of electron microscopic structures ATPase domain ( 40% identity [19]). The two of full-length E. coli DnaB hexamer [21]. Three mono- domains are connected by a linker region (Fig. 1A). In mers, one from each dimer, form the inner rim, and the M. tuberculosis DnaB, MtDnaBn (residues 21–197, other three monomers form the outer rim of the triangle. including the linker) forms a stable dimer in solution The backbones of all six monomers are generally very (Fig. 1C,D). The region spanning residues 135–197 is similar. There are subtle differences in the conforma- required for dimer stability, because a truncated form of tions of C-terminal helices a7 and a8 and in the hairpins the protein (residues 21–134) is monomeric (Fig. 1C,D). connecting them (not shown). These differences indicate MtDnaBn forms a hexamer in the crystallization solu- a small degree of conformational plasticity of the region tion; the crystals of DnaBn contain one slightly asym- closest to the mobile C-terminal domain of DnaB [14]. metric hexamer per asymmetric unit. Three two-fold The interface stabilizing each dimer is formed symmetrical dimers of DnaBn are assembled into a hex- between the C-terminal helix–hairpin–helix (HhH) amer that has approximate three-fold rotational symme- regions of DnaBn (residues 128–165) that protrude try (Fig. 2A). The two-fold and the three-fold symmetry from the globular subdomains (residues 21–127) axes are parallel to each other. The hexamer forms an (Fig. 2B). This explains our observation of the mono- opening in the shape of an equilateral triangle with a meric state of truncated DnaBn (residues 21–134). This side length of 51 A˚. Therefore, the ring can accommo- interface is composed of conserved interdigitating date a cylinder with a maximum diameter of 34 A˚, large hydrophobic residues (Figs 1B and 2B) that bury enough for duplex and even triplex DNA. Indeed, DnaB 2200 A˚2 of solvent-accessible surface area [22]. The is able to encircle and actively translocate on duplex Ca atoms of two glycine residues, Gly136 and Gly143, DNA [20]. The triangular shape is roughly similar to contribute to this interface and allow the helices

Fig. 2. The structure of the MtDnaBn hex- amer. (A) The hexameric ring of MtDnaBn is shown as a ribbon diagram. (B) The HhH–HhH interface. (C) The dimer–dimer interface. In (B) and (C), the residues that form the respective interfaces (see also Fig. 1B) are shown as sticks, with a subset of them labeled.

3066 FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS T. Biswas and O. V. Tsodikov Mycobacterium tuberculosis DnaB helicase to pack snugly against each other. DnaBs of other an inner monomer and the HhH subdomain of an outer bacteria contain either a Gly or an Ala at these monomer. Mutations in this interface significantly two positions. Previously reported mutations in reduce the helicase activity of the DnaB homolog G40P G. stearothermophilus [23] and Salmonella typhimurium [12]. The residues in the dimer–dimer interface are [24] DnaB that cause a priming defect (Val133 and generally, but not universally, conserved (Fig. 1B). Val139 in MtDnaB) lie in this hydrophobic interface. The inner surface of the DnaBn ring contains three These substitutions are likely to abolish DnaB–DnaG positively charged patches, each formed by Arg90, binding by disrupting the structure of DnaBn. The sol- Arg91, Arg 95, and Arg96 (Fig. 3A), that may bind vent-exposed residues of the HhH are generally not the lagging DNA strand during replication. These argi- conserved. The entire HhH region is required for nines are conserved in mycobacteria, but not in other dimerization, because the slightly shorter DnaBn of bacteria (Fig. 1B). Bailey et al. [14] proposed that E. coli (residues 1–161, homologous to residues 1–159 ssDNA binds inside the ring at a different site of MtDnaB[25]) is monomeric. Indeed, residues Ile163 (Fig. 3C), formed by side chains of Arg116, Arg117 and Tyr164 of MtDnaBn, close to the end of the struc- and Arg120 of GsDnaB. Among these residues, only tured region, are located in the dimerization interface Arg116 is conserved (Fig. 1B). The lagging DNA (Fig. 1B). The linker (residues 167–197) is not found in strand must be positioned inside the DnaB ring some- the electron density, consistent with its conformational what differently in different prokaryotes. Such differ- flexibility [14]. This flexibility accommodates large ential positioning is likely to be related to concomitant movements of the C-terminal catalytic domains relative differences in the DnaB–DnaG interface (see the next to each other and to the N-terminal tier [14]. section). The surface residues of DnaB facing the The interface between the dimers is much less exten- center of the ring and the N-terminal side are dramati- sive (1020 A˚2 of buried surface area), which explains the cally different for MtDnaB and GsDnaB (Fig. 3A,C). lower stability of dimer–dimer interactions (Fig. 2C). It In contrast, the surface facing the ATPase domain is is formed predominantly by a set of hydrophobic con- highly conserved and uniformly negatively charged tacts between the two globular subdomains. In addition, (Fig. 3B,D). This conservation must preserve the this interface contains two salt bridges, Arg90–Glu125 nature of interactions between the N-terminal and the and Asp89–Lys126, between the globular subdomain of C-terminal domains of DnaB in different bacteria.

Fig. 3. The electrostatic surface of MtDnaBn and GsDnaBn. The positive sur- face is shown in blue and the negative sur- face in red. The sides facing the C-terminal domain (B, D) in MtDnaBn (A, B) (this study) and GsDnaBn rings (C, D) [14] are similar, whereas the N-terminal faces (A, C) are drastically different. The proposed binding sites for the lagging DNA strand are indi- cated by the arrows.

FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS 3067 Mycobacterium tuberculosis DnaB helicase T. Biswas and O. V. Tsodikov

DnaGc also binds across the dimer–dimer interface of A model of the helicase–primase complex DnaBn and comparably buries 2500 A˚of surface area. DnaBn and its binding partner, DnaGc, are structur- However, GsDnaGc is bound to a different surface of ally similar [10,14,16,26–29]. Therefore, we considered the GsDnaBn ring (Figs 4 and 1B). In an elegant study, the possibility that DnaBn from one hexamer would Chang & Marians [31] demonstrated that the E. coli mimic binding of DnaGc to another hexamer of DnaB mutations Glu32 fi Lys (Ala29 in M. tuber- DnaBn in the crystal lattice. Indeed, in our crystal culosis and either a Glu or an Ala in other bacteria) structure, three DnaBn monomers are bound to the and Tyr105 fi Ala (Tyr102 in M. tuberculosis; almost N-terminal face of each hexameric ring (Fig. 4). This universally conserved) increase the Okazaki fragment 6 : 3 DnaB–DnaG stoichiometry was observed in solu- size by destabilizing the DnaB–DnaG association. No tion [13]. Binding of each MtDnaBn monomer (DnaGc effect was observed for the Glu32 fi Ala mutation. In mimic) to the hexameric ring buries 1900 A˚of solvent- a different system, GsDnaB mutation Tyr88 fi Ala accessible surface area. This interface is different from (Tyr105 in E. coli) was disruptive to the DnaB–DnaG those observed in dimeric structures of the globular complex, whereas the Glu15 fi Ala (Glu32 in E. coli) subdomain of E. coli DnaBn [27,28]. The binding mutation was not [23]. These mutagenesis studies are in surface of the proposed DnaGc mimic involves the excellent agreement with our model of the DnaB– C-terminal helix (the DnaG counterpart of helix a8in DnaG complex: both of these residues are located DnaBn). DnaG mutations in this helix were shown to directly in the putative DnaB–DnaG interface (Figs 4 be disruptive to the DnaB–DnaG complex [30]. Each and 1B). Intriguingly, these two DnaB residues do not DnaGc mimic is bound to the globular subdomains of make direct contacts with DnaG in the GsDnaB– inner and outer monomers across the dimer–dimer GsDnaG structure [14] (see also Fig. 4). interface, thereby stabilizing the hexameric ring of The two distinct positions of DnaG on the DnaB ring DnaB. Such bidentate interaction is consistent with the need not be mutually exclusive. E. coli DnaG priming role of DnaG in contributing to the stability of the activity was demonstrated to be distributive because of hexamer of DnaB and stimulating its activity [5]. the relatively weak association of DnaG with DnaB. N-terminal truncations of the globular subdomain of More than one position of DnaG on DnaB during repli- DnaBn were reported to abolish helicase–primase inter- cation (e.g. in the idle and in the actively priming states) actions [12], in agreement with our structural model. In would be consistent with these observations. Alterna- the structure of the GsDnaB–GsDnaGc complex [14], tively, DnaGc may bind DnaBn somewhat differently in different bacteria. This possibility is strengthened by poor conservation of DnaGc [29] and the N-terminal face residues of the DnaBn ring (Fig. 3A,C). Testing these structural and mechanistic proposals is a subject of future research in this laboratory.

Conclusions In summary, we have obtained a high-resolution struc- ture of the hexameric ring formed by the MtDnaBn. The stable dimers of DnaBn have a propensity for trimerization in the absence of DnaGc. The structure provides the highest-resolution view of the interactions essential for hexamer stability, and suggests a model for the DnaB–DnaG complex that is consistent with previous mutagenesis studies.

Experimental procedures

Fig. 4. The model of the DnaB–DnaG complex. MtDnaBn mono- Cloning, protein expression, and purification mers that mimic DnaGc are shown as the semitransparent gray surface. GsDnaGc (in green) from the GsDnaB–GsDnaGc complex The DnaB constructs (locus tag Rv0058; residues 21–197 and [14] is shown for comparison. The DnaB residues critical for prim- residues 21–134) from Mycobacterium tuberculosis H37Rv ing (Ala29 and Tyr102; see text) are shown as sticks. were amplified by PCR from genomic DNA, and ligated

3068 FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS T. Biswas and O. V. Tsodikov Mycobacterium tuberculosis DnaB helicase between the NdeI and XhoI sites of the modified pET19b Table 1. Crystallographic data collection and structure refinement (EMD Biosciences, San Diego, CA, USA) with an N-termi- statistics. nal decahistidine tag separated from the gene by the Prescis- Data collection sion protease (GE Healthcare, Piscataway, NJ, USA) Space group P31 cleavage site. The proteins were expressed in BL21(DE3) Cell dimensions E. coli cells that were initially grown in LB supplemented a ⁄ b ⁄ c (A˚ ) 114.9 ⁄ 114.9 ⁄ 72.4 )1 with 100 lgÆmL ampicillin at 37 C until the culture a ⁄ b ⁄ c ()90⁄ 90 ⁄ 120 reached an attenuance (D) at 600 nm of 0.4–0.6. Then, the Composition Six monomers per culture was induced with 0.5 mm isopropyl thio-b-d-galacto- asymmetric unit a side at 19 C overnight. The cells were harvested by centrifu- Resolution 50.0–2.0 (2.06–2.00) gation at 5000 g for 10 min, resuspended, and lysed by Rmerge 0.083 (0.64) sonication in lysis buffer (40 mm Tris, pH 8.0, 400 mm NaCl, I ⁄ rI 12.2 (2.0) Completeness 0.99 (0.95) 10% glycerol, 2 mm b-mercaptoethanol). The clarified lysate Redundancy 5.5 (4.1) was passed through an IMAC (GE Healthcare) Ni2+ col- Refinement umn, following the manufacturer’s instructions. The tag was Resolution (A˚ ) 30.0–2.0 (2.05–2.00) cleaved with Prescission protease (GE Healthcare) overnight No. of reflections 67 873 (3634) at 4 C. The protein was then concentrated using an Amicon R ⁄ Rfree 0.23 ⁄ 0.28 (0.30 ⁄ 0.35) Ultra-15 centrifugal filter unit (Millipore, Bellerica, MA, No. of atoms USA) and purified further on an S-200 size exclusion column Protein 6429 (GE Healthcare) equilibrated in 40 mm Tris, pH 8.0, 200 mm Water 421 NaCl, 2 mm b-mercaptoethanol and 0.1 mm EDTA. Frac- rmsd ˚ tions containing DnaBn were pooled, and concentrated to Bond lengths (A) 0.011 ) Bond angles ( ) 1.23 10 mgÆmL 1 using an Amicon Ultra-15 centrifugal filter unit. The concentrated protein was used for crystallization. a The highest-resolution-shell values are shown in parentheses.

Protein crystallization, data collection, and structure determination and refinement Acknowledgements

Initial high-throughput crystallization screening of We thank Dr Nick Dixon for helpful discussions on MtDnaBn was performed at the Hauptman–Woodward the manuscript. Institute [32]. Single crystals of size 0.1 · 0.1 · 0.1 mm were The research was supported by start-up funds to grown in 4–6 weeks by microbatch crystallization in mineral O. V. Tsodikov from the College of Pharmacy at the oil at 21 C. The drops contained a mixture of 1 lL of the University of Michigan. We would like to thank the concentrated protein (see above) and 1 lL of crystallization staff of LS-CAT, and especially Spencer Anderson, for solution. Crystallization solution consisted of 100 mm citrate assistance with the data collection at the APS. sodium salt (we used a stock solution of 1 m citric acid adjusted to pH 4.0 with 1 m sodium citrate), 100 mm dibasic References potassium phosphate, and 17.5% poly(ethylene glycol) 8000. The crystals were gradually transferred to crystallization 1 Hirota Y, Ryter A & Jacob F (1968) Thermosensitive solution with 16.5% glycerol, incubated overnight, and then mutants of E. coli affected in the processes of DNA flash frozen in liquid nitrogen. X-ray diffraction data were synthesis and cellular division. Cold Spring Harb Symp collected at 100 K at LS-CAT beamline (sector 21-ID) at the Quant Biol 33, 677–693. Advanced Photon Source at the Argonne National Labora- 2 LeBowitz JH & McMacken R (1986) The Escherichia tory. The diffraction data were processed using the hkl2000 coli dnaB replication protein is a DNA helicase. J Biol suite [33]. The crystals were not merohedrally twinned [34]. Chem 261, 4738–4748. The structure of MtDnaBn was determined by molecular 3 Baker TA, Sekimizu K, Funnell BE & Kornberg A replacement using program phaser [35], with residues 5–112 (1986) Extensive unwinding of the plasmid template of the monomeric DnaB from Thermus aquaticus [10] as a during staged enzymatic initiation of DNA replication search model. The missing part of the structure was then from the origin of the Escherichia coli chromosome. built into the unambiguous Fo–Fc electron density. The struc- Cell 45, 53–64. ture of MtDnaBn was iteratively rebuilt and refined by using 4 Schaeffer PM, Headlam MJ & Dixon NE (2005) Pro- programs coot [36] and refmac [37], respectively, to a final tein–protein interactions in the eubacterial . resolution of 2.0 A˚. The data collection and refinement statis- IUBMB Life 57, 5–12. tics are given in Table 1. The structure and the diffraction 5 Corn JE & Berger JM (2006) Regulation of bacterial data were deposited in the (code: 2r5u). priming and daughter strand synthesis through

FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS 3069 Mycobacterium tuberculosis DnaB helicase T. Biswas and O. V. Tsodikov

helicase–primase interactions. Nucleic Acids Res 34, interacts with a linker region that joins the N- and 4082–4088. C-domains of DnaB and induces the formation of 3-fold 6 Kaplan DL (2000) The 3¢-tail of a forked-duplex steri- symmetric rings. Nucleic Acids Res 32, 2977–2986. cally determines whether one or two DNA strands pass 19 Leipe DD, Aravind L, Grishin NV & Koonin EV through the central channel of a replication-fork heli- (2000) The bacterial replicative helicase DnaB evolved case. J Mol Biol 301, 285–299. from a RecA duplication. Genome Res 10, 5–16. 7 Galletto R, Jezewska MJ & Bujalowski W (2004) 20 Kaplan DL & O’Donnell M (2002) DnaB drives DNA Unzipping mechanism of the double-stranded DNA branch migration and dislodges proteins while encircling unwinding by a hexameric helicase: quantitative two DNA strands. Mol Cell 10, 647–657. analysis of the rate of the dsDNA unwinding, proces- 21 San Martin MC, Stamford NP, Dammerova N, Dixon sivity and kinetic step-size of the Escherichia coli DnaB NE & Carazo JM (1995) A structural model for the helicase using rapid quench-flow method. J Mol Biol Escherichia coli DnaB helicase based on electron 343, 83–99. microscopy data. J Struct Biol 114, 167–176. 8 Mok M & Marians KJ (1987) The Escherichia coli prep- 22 Tsodikov OV, Record MT Jr & Sergeev YV (2002) rimosome and DNA B helicase can form replication Novel computer program for fast exact calculation of forks that move at the same rate. J Biol Chem 262, accessible and molecular surface areas and average 16644–16654. surface curvature. J Comput Chem 23, 600–609. 9 Tanner NA, Hamdan SM, Jergic S, Schaeffer PM, 23 Thirlway J & Soultanas P (2006) In the Bacillus Dixon NE & van Oijen AM (2008) Single-molecule stearothermophilus DnaB–DnaG complex, the activities studies of fork dynamics in Escherichia coli DNA repli- of the two proteins are modulated by distinct but over- cation. Nat Struct Mol Biol 15, 170–176. lapping networks of residues. J Bacteriol 188, 10 Bailey S, Eliason WK & Steitz TA (2007) The crystal 1534–1539. structure of the Thermus aquaticus DnaB helicase 24 Stordal L & Maurer R (1996) Defect in general priming monomer. Nucleic Acids Res 35, 4728–4736. conferred by linker region mutants of Escherichia coli 11 Biswas SB, Chen PH & Biswas EE (1994) Structure and dnaB. J Bacteriol 178, 4620–4627. function of Escherichia coli DnaB protein: role of the 25 Miles CS, Weigelt J, Stamford NP, Dammerova N, N-terminal domain in helicase activity. Biochemistry 33, Otting G & Dixon NE (1997) Precise limits of the 11307–11314. N-terminal domain of DnaB helicase determined by 12 Wang G, Klein MG, Tokonzaba E, Zhang Y, Holden NMR spectroscopy. Biochem Biophys Res Commun 231, LG & Chen XS (2008) The structure of a DnaB-family 126–130. replicative helicase and its interactions with primase. 26 Syson K, Thirlway J, Hounslow AM, Soultanas P & Nat Struct Mol Biol 15, 94–100. Waltho JP (2005) Solution structure of the helicase- 13 Bird LE, Pan H, Soultanas P & Wigley DB (2000) interaction domain of the primase DnaG: a model for Mapping protein–protein interactions within a stable helicase activation. Structure 13, 609–616. complex of DNA primase and DnaB helicase from 27 Fass D, Bogden CE & Berger JM (1999) Crystal Bacillus stearothermophilus. Biochemistry 39, 171–182. structure of the N-terminal domain of the DnaB hexa- 14 Bailey S, Eliason WK & Steitz TA (2007) Structure of meric helicase. Structure 7, 691–698. hexameric DnaB helicase and its complex with a 28 Weigelt J, Brown SE, Miles CS, Dixon NE & Otting G domain of DnaG primase. Science 318, 459–463. (1999) NMR structure of the N-terminal domain of 15 Singleton MR, Sawaya MR, Ellenberger T & Wigley E. coli DnaB helicase: implications for structure rear- DB (2000) Crystal structure of T7 gene 4 ring helicase rangements in the helicase hexamer. Structure 7, indicates a mechanism for sequential hydrolysis of 681–690. nucleotides. Cell 101, 589–600. 29 Su XC, Schaeffer PM, Loscha KV, Gan PH, Dixon NE 16 Oakley AJ, Loscha KV, Schaeffer PM, Liepinsh E, & Otting G (2006) Monomeric solution structure of the Pintacuda G, Wilce MC, Otting G & Dixon NE (2005) helicase-binding domain of Escherichia coli DnaG prim- Crystal and solution structures of the helicase-binding ase. FEBS J 273, 4997–5009. domain of Escherichia coli primase. J Biol Chem 280, 30 Tougu K & Marians KJ (1996) The extreme C terminus 11495–11504. of primase is required for interaction with DnaB at the 17 Lu YB, Ratnakar PV, Mohanty BK & Bastia D (1996) replication fork. J Biol Chem 271, 21391–21397. Direct physical interaction between DnaG primase and 31 Chang P & Marians KJ (2000) Identification of a region DnaB helicase of Escherichia coli is necessary for opti- of Escherichia coli DnaB required for functional inter- mal synthesis of primer RNA. Proc Natl Acad Sci USA action with DnaG at the replication fork. J Biol Chem 93, 12902–12907. 275, 26187–26195. 18 Thirlway J, Turner IJ, Gibson CT, Gardiner L, Brady 32 Luft JR, Collins RJ, Fehrman NA, Lauricella AM, K, Allen S, Roberts CJ & Soultanas P (2004) DnaG Veatch CK & DeTitta GT (2003) A deliberate

3070 FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS T. Biswas and O. V. Tsodikov Mycobacterium tuberculosis DnaB helicase

approach to screening for initial crystallization crystallographic software. J Appl Crystallogr 40, 658– conditions of biological macromolecules. J Struct Biol 674. 142, 170–179. 36 Emsley P & Cowtan K (2004) Coot: model-building 33 Otwinowski Z & Minor W (1997) Processing of X-ray tools for molecular graphics. Acta Crystallogr D Biol diffraction data collected in oscillation mode. Macromo- Crystallogr 60, 2126–2132. lec Crystallogr A 276, 307–326. 37 Murshudov GN, Vagin AA & Dodson EJ (1997) 34 Yeates TO (1997) Detecting and overcoming crystal Refinement of macromolecular structures by the maxi- twinning. Methods Enzymol 276, 344–358. mum-likelihood method. Acta Crystallogr D Biol Crys- 35 Mccoy AJ, Grosse-Kunstleve RW, Adams PD, tallogr 53, 240–255. Winn MD, Storoni LC & Read RJ (2007) Phaser

FEBS Journal 275 (2008) 3064–3071 ª 2008 The Authors Journal compilation ª 2008 FEBS 3071